Re: Tj'a-ts'añ and Tech (was Georgian and Arabic N-A word order)

At 22:32 16/06/00 CDT, you wrote:
>>From: Christophe Grandsire <Christophe.Grandsire@...>
>
>>Interesting. Looks a little like my Tj'a-ts'a~n which has all four cases
>>'ergative', 'absolutive', 'nominative' and 'accusative' (plus a dative used
>>also as the unmarked case for copular sentences) depending on the relation
>>set by the verb between the subject and the object and what is the status
>>of each one. Really a kind of active marking I think but more complex.
>
>I'm probably only going to have three cases governing the subject and the
>verb.  Georgian ergativity has little semantic role; it's mostly formal.
>Tech will probably use the three declensions applicable to subjects.  I
>don't see much *need* for four cases, but there may be natlangs that
>actually do that, at least in a formal sense...
>
>The main issue I'll be wrestling with is how to casemark entities in a
>causative verb construction.  How do natlangs decline nouns referring to who
>causes the action, who actually carries out the action, and who is the
>recipient of the action...  (I'm leading toward this: 'My dear A-vocative,
>B-ergative of/from C-genitive caused D-nominative to <verb> E-accusative to
>F-dative with/by means of G-instrumental in a H-ablative way at/in
>I-locative".)
>
I don't know for natlangs, but Tj'a-ts'a~n solves the problem by an
optional slot in the verbal complex, which looks like the subject and
object slots (it agrees with the causative noun in gender and case), but
has also a "causative morpheme" in it, so any case can be causative,
depending on the meaning of the causative relation. Generally, the case
employed for the causative is of higher volitionality than the case used
for the subject, but it's not always the case, especially when the
causative noun is inanimate.

Anyway, I don't think natlangs use this. Tj'a-ts'a~n is an alien language
and has quite a few alien features (even if I find them naturalistic) like
the fact that the genitive is not a case but a special category (nominal
categories are: noun, genitive, attributive and adjective, the last three
used to complete the first one).

>That's a sentence using all nine of my cases (what I got so far), but I
>expect a tenth to surface somehow.  Theoretically, if all the nouns are
>pronouns, that could theoretically be ALL ONE WORD!!  (Sheesh, that's gonna
>be right ugly...)
>
Why? It all depends on how the pronouns look like :) .

>>In Tj'a-ts'a~n, the core cases are basically used this way:
>>ergative: animate volitional agent
>>nominative: animate unvolitional agent or inanimate agent
>>absolutive: non-agent, non-patient
>>accusative: patient
>
>So, in effect, ergative has an active role in the action, while nominative
>is passively involved in the action.  Not the same thing as
>active/passive/middle voice, but similar.  But your absolutive is a
>non-agent, non-patient -- so what would that be?
>
Like the experiencer of verbs like 'to see' I would say, but I'm not
completely sure. It's more a case used when you cannot really decide
whether something originates or undergo the action. The use of cases in
Tj'a-ts'a~n is quite complex (and it gets even more complex when you begin
to use the local cases for subjects and/or objects - which is perfectly
possible - :) ).

>>I use 'agent' and 'patient' because in Tj'a-ts'a~n it's not the cases that
>>show what is subject and what is object but the agreement with the verb (in
>>the verbal complex, there are two 'slots' for the subject and the object
>>where agreement affixes in gender at least and even case are put). Cases
>>show actual functions, thus a noun in a local case can be subject or object
>>of a verb!
>
>Oh, so you got an active language.  Georgian is kind of an active language,
>but an unorthodox one.  It's kinda on the border of being ergative and being
>active but it still has a lot of nominative-accusative features.  (Modern
>Georgian, or at least the standard T'bilisi dialect, uses the dative case
>for accusatives as well as locatives.  Tech uses three distinct cases, as I
>said above.)
>
>I was also thinking of the 'trigger system' of Philippine languages, but I'm
>not sure that's what you have...
>
I thought about that but in trigger languages the form of the verb says the
role of the trigger, which has always the same mark, whatever its real
function, whereas in this case there is no uniform trigger mark, more a
kind of agreement in case, and the cases have a semantic role more than a
functional one.

                                                Christophe Grandsire
                                                |Sela Jemufan Atlinan C.G.

"Reality is just another point of view."

homepage : http://rainbow.conlang.free.fr
(ou : http://www.bde.espci.fr/homepages/Christophe.Grandsire/index.html)